My Skills
My Experience
Student Research Assistant (HiWi)
NHR@FAU (Erlangen National High Performance Computing Center)
- Building and validating data pipelines that process parallel MPI execution traces for SP3 of DFG-funded Mod4Comp research unit (FOR 5880)
- Supporting pre-execution energy forecasting for massively parallel HPC workloads under supervision of Dr. Ayesha Afzal
Software Engineer
Dubizzle Labs (Bayut · OLX · Zameen)
- Elasticsearch KPI dashboards over 12M+ records with sub-second query performance using nested aggregation pipelines
- AI analytics modules via OpenAI APIs in Jarvis CRM, eliminating manual performance summaries across 3 business units
- Real-time GPS movement tracking system improving field task tracking accuracy by 90%
- Automated revenue workflows integrating Slack, Stripe, Tabby, Tamara, and Exchange Rate APIs
Associate Software Engineer
Dubizzle Labs
- Built Ingress, an ML-powered microservice processing 10,000+ daily sales calls using RabbitMQ, OpenAI, and custom ML models
- Automated finance reconciliation syncing 1.7B+ PKR in Oracle NetSuite transactions with zero manual intervention
- Integrated Checkout.com with full 3D-secure compliance in Polaris CRM
- Migrated 50,000+ users to Bayut.sa with zero data integrity issues
Machine Learning Intern
KICS, UET Lahore
- Built CNN pipeline for impulsive acoustic event detection using 2D spectrograms
- Configured vSphere dual-OS server with VMRC remote access for multi-researcher ML environment
About Me
I'm a computer scientist and MSc Data Science student at FAU Erlangen-Nürnberg, Germany. My work sits at the intersection of HPC systems, energy-efficient parallel computing, and applied machine learning. At NHR@FAU, I work on simulator frameworks for runtime and energy prediction of massively parallel MPI programs as part of the DFG-funded Mod4Comp research unit. Before academia, I spent 2+ years at Dubizzle Labs building production systems — Elasticsearch pipelines, ML microservices, and finance automation at scale.
My Education
M.Sc. Data Science
FAU Erlangen-Nürnberg · Erlangen, Germany
B.Sc. Computer Science
COMSATS University Islamabad · Lahore, Pakistan
Thesis: Deep Learning-Based Approach for Creation of Synthetic Videos
Courses
Image Processing with Python Track
07/2022DataCamp
Object Detection & Biomedical Image Analysis using Python and Keras
Python Fundamentals Track
07/2022DataCamp
Work Authorization
German student residence permit (§16b), valid for 120 full / 240 half days.
Languages
English (C1 — Professional Working Proficiency), Urdu (Native)
My Projects
Mod4Comp SP3 — MPI Energy Simulator
Simulator framework for runtime and energy prediction of massively parallel MPI programs. Built data pipelines processing parallel execution traces for pre-execution HPC energy forecasting.
Puppet Master — Synthetic Video Generation
Deep learning system for high-resolution synthetic video generation, ranked top 3 among submissions. Combined motion transfer, Real-ESRGAN upscaling, image-to-image translation, and facial landmark extraction on VGGFace2HQ and VoxCeleb2 datasets, served via a Flask backend for real-time generation.
Ruby-Based File-Aware GPT-4 Assistant
Ruby-based file-aware GPT-4 assistant built on OpenAI's Assistants API v2, with custom Faraday HTTP integrations for endpoints not yet in the official Ruby gem. Handles file uploads, assistant configuration, thread/message handling, run automation, and structured JSON output parsing.
Ingress — ML Call Transcription Microservice
ML-powered microservice handling 10,000+ daily multilingual sales calls using RabbitMQ, OpenAI Whisper, and custom models. Enabled automated quality scoring for support teams.
Elasticsearch KPI Analytics Platform
Real-time sales dashboards with drill-downs, histograms, and trajectory views over 12M+ records. Sub-second query performance via nested Elasticsearch aggregation pipelines.